In today`s world where a lot of mobile internet enabled devices are emerging,
there needs to be some mechanism that can support the mobility of devices
from one place to another while they are communicating. This mobility
in a WLAN environment can be of two types
i.e., Intra-domain mobility and Inter-domain mobility (Li, 2004). A lot
of techniques have been proposed to handle Intra-domain mobility (Hsich
et al., 2003; Ramjee et al., 2000; András, 1999)
and Inter-domain mobility (Buddhiket et al., 2003; Khan et al.,
2003; Perkins, 2002). Intra-domain mobility protocols emphasize on reducing
the handoff latency and eventually minimizing the data loss during the
handoff, as the mobility is local. In Inter-domain mobility, where long
registration/handoff delay is still an issue, are widely left to Mobile
There needs to be a mechanism to transfer the transient data to the new
location in a minimum amount of time backed by a buffering mechanism to
minimize the data loss.
Mobile IP being an Internet Engineering Task Force (IETF) standard communications
protocol allows mobile device users to move from one network to another
while maintaining their permanent IP address. This is achieved by having
a permanent home IP address and temporary Care of Address (CoA) acquired
from the foreign network (Perkins, 2002). The reason for using two IP
addresses per host is because whenever a mobile device with one IP address
initiates a session and after that changes IP address, the ongoing sessions
tear down. It results from the fact that the IP address is overloaded
with two functions i.e., host identification by TCP layer and network
attachment information for routing purposes (Bhagwat et al., 1996).
A hierarchical design approach has been proposed to optimize mobility
in Mobile IP networks by Perkins and Johnson (2001) in their internet-draft
document, while a mechanism to eliminate transient data losses by buffering
at the Base Station Foreign Agents has also been presented (Perkins and
Wang, 1999). The old Base Station Foreign Agent maintains buffers for
every MN and if MN changes its Base Station Foreign Agent, the old one
sends this buffer to the new Base Station Foreign Agent. A similar approach
using Route Optimization has been utilized by Erom et al. (2002).
Mobile NAT is a mobility management scheme that incorporates Network
Address Translator (NAT) to provide transparent mobility and also to minimize
the usage of public IP address space for Ipv4 by using private IP addresses
(Buddhikot et al., 2003). Similar to Mobile IP, it also uses two
IP addresses to locate the mobile device. Whenever a mobile device moves
from one administrative domain to another during a session with a correspondent
node (e.g., a web server), an entity called Mobility Manager in the visited
domain contacts the Mobility Manager of the home domain for that particular
session to get the session parameters and informing the home AN to forward
the packets arriving from the correspondent node to the new AN through
a tunnel. This technique however does not handle the problem of transient
data arriving from the home NAT domain to the previous NATed domain.
Another buffering scheme by Khan et al. (2003) proposes a buffering
method in which the transient data is buffered for each MN on the Foreign
Agent for a specific period of time, called the absence detection interval.
If the MN is found absent from the Foreign Agent coverage area, even after
the absence detection interval expires, then all the data destined for
the MN is sent back to its Home Agent (HA). HA on receiving the MN`s new
registration, tunnels the data to the new location in addition to the
data arriving at the HA from the Correspondent Node (CN). Problem with
this technique is that it only buffers the data after it detects the MN
absence from the Foreign Agent coverage region thus resulting in loss
of data. This scheme can also increase the overall handoff time if the
home domain for a mobile node is far away from the previous Foreign Agent.
It is, therefore, intended to devise a method by which we only need to
buffer data for those MN that can possibly leave the current domain and
enter a neighboring domain, greatly improving the efficiency. By using
the access points at the boundaries of a domain we can speedup the process
of transferring the transient data to the new location. We have utilized
the Access points present at the boundary of a domain to transfer the
data to the new location in minimum time.
The study is organized in a way that the Proposed scheme is described
in detail with the help of figures and tables and also its comparison
with existing schemes is shown. Enhancement of the proposed scheme is
The proposed scheme is based on the idea of placing access points at
those sections of a domain from where a MN can possibly enter the boundary
of the neighboring domain. These access points are a new component called
the Boundary Access Point (BAP) as shown in Fig. 1.
In addition to the introduction of BAP we have also incorporated the
capability of buffering at the Home agent or gateway router of a domain
and on the Foreign Agents thus helping in preserving the transient data
during a handoff process. The key concepts and the roles of different
components in the architecture are as below.
Tasks performed by a Boundary Access Point (BAP): BAP is responsible
for immediately informing the home or gateway router about any mobile
device entering or leaving its coverage area in addition to forwarding
the dummy packets to the paging server. Every BAP has a unique BAP ID
thus helping the gateway router in finding out the neighbor domain where
a MN could possibly be heading.
||BAP in a domain
||MN entering a BAP coverage area
Whenever a MN enters a BAP coverage area it receives the subnet or domain
information including the BAP id thus keeping the MN informed about any
subnet or domain changes. The above discussion is represented in the Fig. 2
Paging server: The architecture also contains a Paging Server
(PS) to maintain up-to-date location information about a MN. Every MN
sends a location update packet whether it has any sessions or not as shown
in Fig. 3.
The frequency of sending the dummy packet depends on the presence or
absence of a session that is if a MN has sessions then it will send the
dummy packet with a higher frequency otherwise with a low frequency.
||MN sending a dummy packet under a simple AP
The PS can be a separate entity or can be incorporated in a home agent
or gateway router.
Tasks performed by the home agent: The Home agent of any domain
would have knowledge about its neighbor domains so that when a MN leaves
the domain, the neighbor domain agents can be contacted for transferring
the transient data.
Every Home agent has the knowledge about a MN present under a specific
BAP having any sessions. Based on this information, the HA buffers packets
for those MN. The data present in the buffer gets overwritten when the
HA does not receive any information from the BAP about the MN exit for
a time duration equal to the reception of the dummy packet.
If a MN exits the BAP coverage area, the HA gets informed about that
by the BAP. It immediately stops overwriting the buffered data and sends
request for the MN to the PS as shown in Fig. 4.
If the information present with the PS regarding that specific MN is
old, the HA will send a tunnel establishment request to the neighbor domain
agent so that a copy of the buffered data can be transferred in addition
to transferring the transient data as presented in Fig. 5.
If the neighbor domain also does not have the info about the MN when
the packets arrive through the tunnel, it would drop the packets and it
could easily be assumed that the MN is switched off.
Tasks performed by the foreign agent: When the MN enters a FA
domain, it gets a new Care of Address (CoA) and informs about its Home
ID. The Home ID helps the FA in contacting with the HA when the MN exits
the FA domain and enters a new domain.
||Steps after MN exit
||Tunnel establishment phase
||HA gets informed about FA2
shows the steps when a MN exits the FA domain.
After MN node exit, the FA performs the same steps that are performed by
the HA. In addition to that, FA1 also informs the HA about the new domain
so that the transient data can be directed to the new location. When the
HA gets the new domain info, it creates a tunnel with the new domain.
After that the transient data starts flowing to the new FA2. If the FA2
has info about that MN, it will deliver the data to it, otherwise informs
the HA that the MN has not entered in its domain after dropping the packets.
USABILITY OF BAP SCHEME WITH OTHER SCHEMES
BAP scheme can be used to make the other inter-domain mobility schemes
perform even better. The BAP scheme can easily be mapped to the Mobile
Ipv4 as it contains the same components as the HA and the FA. The proactive
buffering scheme can also get benefit by buffering for only those MN that
is present under the BAP. Mobile NAT can also use the same idea to improve
the handoff process in addition to having a NAT-ed domain concept.
ENHANCEMENT TO THE PROPOSED SCHEME
Buffering for only those MN that are present under a BAP and are expected
to move to a new domain can further enhance the proposed scheme. As there
may be some MN that would not move at all during an established sessions
having present under a BAP, we need not to buffer packets for them. We
shall only buffer data for those MN that are present under BAP and will
most probably move to a new domain.
Incorporating a profiling or a configuration software approach that can
define a MN session as Static session or a Mobile session can achieve
COMPARISON WITH SEQUENCE DIAGRAMS
Figure 7 shows a normal MIPv4 approach that does not
provide any buffering capability. In this approach the transient data
gets dropped until a new binding update is not received.
Figure 8 shows the sequence diagram of Mobile NAT scheme
that clearly shows that this scheme also does not provide any buffering
||MIP flow chart
||Mobile NAT flow chart
As a MN uses Physical IP (PIP) and a Virtual IP (VIP) so after entering
a new NAT-ed domain the MN sends a request to the DHCP server so that it
can allot the requested IP`s. Figure 9
shows the sequence
diagram of the Proactive buffering approach. This approach has the problem
of dropping the transient data before detecting that the MN has exited its
domain as shown.
The other problem with this scheme is that when a MN moves to a new domain
and the Absence Detection Interval expires, the old FA needs to send the
buffered and the transient data back to the HA without considering the
fact that the HA can be at a distant location. This introduces delay,
thus preventing the early transfer of transient packets to the MN.
||Proactive buffering flow chart
||MN exits from home domain
||IP assignment to MN in new domain
shows the proposed scheme when a MN leaves
the home domain and it can be observed that the transient data is already
being buffered when the MN exit info reached the HA thus reducing the packet
losses. Here when the HA is informed about MN exit by the BAP, it contacts
the PS. If the PS also has information older than the expected one, the
HA will send a tunnel establishment request to the new domain FA and will
send the buffered data.
Figure 11 shows the process when the MN enters a new
domain. When the MN enters new domain, it gets the subnet information
of the new domain and thus sends a request for a new IP address.
The Table 1 below shows a comparison of various schemes
under certain parameters with the proposed scheme:
||Comparison of available schemes with proposed scheme
We have proposed a scheme that can handle the issue of transient data
loss when a MN moves from one administrative domain to the other. This
is achieved by introducing a new component called the Boundary Access
Point (BAP). The BAP can be used to monitor the MN that could possibly
enter a neighbor domain. To support this idea, buffering is done at the
Home Agent (HA) and Foreign Agent (FA) so that buffers can be maintained
for the MN present under BAP coverage area. Whenever a MN moves, a copy
of these buffers is sent to the new location.
Another important aspect of this scheme is the earliest delivery of the
transient data to the new location. This also motivates us to name our
scheme as a Fast inter-domain mobility scheme. All the inter-domain mobility
schemes discussed here except the proactive buffering scheme rely on informing
the previous gateway after the MN moves to the new administrative domain.
This process greatly increases the delay in transferring the transient
data to the new location. Proposed scheme uses BAP to aid in sending buffered
and transient data packets to new location as soon as the MN exits the
previous administrative domain.
Finally, this scheme can provide benefits to other inter-domain mobility
schemes such as Mobile Ipv4, Mobile NAT and proactive buffering schemes.